JP7301443B2 - Methods, apparatus, systems and computer program products related to filtration processes - Google Patents

Description

Filtration processes are used in many industries, such as the biotechnology industry, the pharmaceutical industry, the food industry, etc., for example, to separate various components of a substance. Types of filtration processes include dead-end filtration and cross-flow filtration. In dead-end filtration, the fluid (known as the "feed") is passed through a filter, typically containing a membrane, which traps certain solids and particles while the remaining fluid (the "permeate") known as "liquor" or "filtrate") passes through the filter. In cross-flow filtration (also known as "tangential flow filtration"), the feed liquid is principally propagated tangentially along the surface of the filter. A pressure differential (known as trans-membrane pressure (TMP)) is applied across the filter to create a positive pressure on the feed side relative to the opposite side (known as the permeate side) make it This allows some of the material (material with dimensions smaller than the pore size) to pass through the filter. Fluid that does not pass through the membrane (known as "residue") is typically returned to a supply vessel or other vessel and may be recycled through the membrane.

Filter characteristics are important factors in the design and implementation of such filtration processes. Factors such as pore size and filter size affect the suitability of a filter for a given process, as are factors such as the time required to produce a given quality of desired product.

1 schematically illustrates an exemplary simulation device according to one example; 1 schematically illustrates an exemplary filtering system in which a simulated filtering process according to one example can be implemented; Figure 4 schematically shows the filter device in use; FIG. 4 is a flow diagram illustrating a method for selecting a filter device according to one example; FIG. 1 is a schematic diagram of an example simulation device and filtering system; FIG. 1 is a flow diagram illustrating a method of monitoring a filtration process according to an example;

FIG. 1 illustrates an apparatus in the form of a computer-type device (hereinafter simulation apparatus 100) configured to perform an example simulation process. The simulation device 100 includes at least one processor 102 communicatively coupled to a non-transitory computer-readable storage medium 104 having a set of computer-readable instructions 106 stored thereon. The set of computer readable instructions 106 may be executed by at least one processor to cause at least one processor 102 to perform any example method described herein. Simulation device 100 may also include interface 108 . Interface 108 may comprise a user interface such as a keyboard, mouse, touch screen, other input device, and the like. Alternatively or additionally, interface 108 may comprise a communication interface to communicate remotely with a user or other entity, eg, via wired or wireless communication, via the Internet.

At least one processor 102 may include a microprocessor, microcontroller, processor module or subsystem, programmable integrated circuit, programmable gate array, or other control or computing device. Computer-readable storage medium 104 may be embodied as one or more computer-readable storage media. The computer readable storage medium 104 may include various types of memory, such as semiconductor memory devices such as dynamic random access memory modules and static random access memory modules (DRAM and SRAM), erasable programmable read only memory modules (EPROM), electrical erasers. programmable read-only memory module (EEPROM), flash memory, etc.), magnetic disk (fixed disk, floppy disk, removable disk, etc.), other magnetic media (tape, etc.), optical media (compact disk (CD), DVD, etc.) , may include other types of storage devices. Computer readable instructions 106 may be stored on a single computer readable storage medium or may be stored on multiple computer readable storage media.

FIG. 2 schematically illustrates an exemplary filtering system 200 in which a simulated filtering process can be performed according to one example. The exemplary filtration system 200 of FIG. 2 is for performing a cross-flow filtration process.

Filtration system 200 includes supply vessel 202 and filter device 204 , which includes membrane 206 and filter channel 208 . A fluid (feed liquid) is passed from the supply container 202 to the first end of the filter channel 208 . The feed liquid then flows through filter channels 208 across the surface of filter membrane 202 . The pressure differential across membrane 206 causes some components (filtrate), eg, components having dimensions smaller than the pores of membrane 206 , to pass through membrane 206 toward permeate side 210 of filter device 204 . In the illustrated example, the filtrate flows out of filter device 204 and through outflow tube 212, from where it may be collected, for example. Components (residues) that do not pass through the membrane are returned to supply vessel 202 . The residue can be reused as feed in further filtration cycles.

Examples of filter devices 204 that may be used with the system 200 of FIG. 2 include cartridge filters and cassette filters. In cartridge filters (also known as hollow fiber filters), the membrane comprises a set of parallel hollow fibers. The feed liquid passes through the lumen of the fibers and the filtrate passes through the fibers and can be collected from the outside of the set of fibers. Cassette filters include a housing that holds a plurality of (typically flat) membrane sheets held apart from each other by support screens. The feed liquid passes between the membrane sheets. Filtrate may pass through the sheet and be collected from the side of the sheet opposite the side through which the feed liquid passes.

Filtration processes such as illustrated in FIG. 2 can be used for a wide variety of purposes, such as biotechnology, pharmaceutical, petrochemical, food technology purposes. Biotechnology processes in which the system 200 can be used include cell harvesting processes, cell and lysate sorting processes, protein fractionation processes, concentration processes, and diafiltration processes. In some cases, the purpose of filtration is to obtain a filtrate, i.e., the filtrate is the desired product for further use, while in other cases the desired product is the residue, or the mixture of residue and filtrate. Note that it is both.

FIG. 3 shows an example of filter device 204 in use. Filter device 204 includes filter channel 208 of half height (radius) h through which fluid flows over the surface of filter membrane 206 of length L with velocity u(y). A layer 300 of solidified solute (often referred to as a “cake” or “gel” layer) forms on the surface of membrane 206 and has a thickness δ that can vary along the length of membrane 206 . The filtrate flux J(x) (the volume of filtrate flowing through the membrane 206 per unit area per unit time) can vary along the length of the membrane 206 due to the effect of the gel layer 300, for example.

FIG. 4 is a flow diagram illustrating a method 400 of selecting a filter device for an example filtration process.

At 402 , simulation device 100 receives input data via interface 108 . At 404, simulation device 100 identifies one or more filter characteristics for each of a plurality of filter characteristics for a filtering process, such as the process described above with respect to FIG. 2, based on the received input data.

Examples of filter characteristics include the geometrical characteristics of the filter, e.g., the length of the membrane 206 over which the feed flows, the area of the membrane, the height (diameter) of the channel through which the feed flows over the membrane 206. Alternatively, film characteristics such as half height (radius) and porosity can be mentioned. In one example, the input data comprises values of these properties, ie receives filter properties directly from the user via interface 108, for example. In other examples, the input data identifies multiple candidate filters to the simulation device 100, eg, by providing a filter identifier for each of the multiple candidate filters. In this case, simulation device 100 may retrieve data from a data store that stores the association between filter identifiers and filter characteristics of these filters, for example in the form of a lookup table (not shown). The data store storing the associations can be the computer readable storage medium 104 shown in FIG. 1, or a remote data store accessed via an interface 108, such as a data store on a server device. But it is possible.

At 406, the simulation device 100 identifies the process parameters of the filtration process, ie, the parameters of the filtration process to be simulated. Process parameters include, for example, the type of filtration process (e.g., cross-flow filtration or dead-end filtration), partition coefficient, initial feed volume, initial feed composition, initial feed viscosity, initial feed concentration, initial feed temperature, A combination of initial feed solubility, initial liquid permeability is included. Values for these parameters may be received, for example, from a user via interface 108, or an identifier, process, or process type may be received via interface 108, similar to that described above with respect to filter characteristics, and A parameter is determined by retrieving data based on the identifier.

At 408, simulation device 100 performs a computer simulation process of the filtering process for each of the plurality of candidate filters based on the identified filter characteristics and the identified process parameters. Examples of simulations are described in detail below.

At 410, simulation device 100 determines the output characteristics of the filtering process based on the simulation. For example, the simulation device 100 may determine one or more of the following: filtrate volume, filtrate composition, filtrate viscosity, residue volume, residue composition, residue viscosity, properties related to the formation of a "cake" layer (such as its effect on the liquid permeability of membrane 206). can be determined. Output characteristics can be output as static quantities (eg, output quantities after a period of time). Alternatively or additionally, the output may represent time-varying of one or more output characteristics. For example, one or more output characteristics may be represented as a graph showing the variation of that characteristic over time.

At 412, one of the plurality of candidate filters is selected based on one or more output characteristics. In some examples, simulation device 100 makes a selection based on the determined output characteristics and provides an indication of the selection to the user, e.g. can be displayed. For example, simulation device 100 compares the determined output characteristic to a desired output characteristic (which may, for example, be input by a user via interface 108) and finds a filter that is closer to the desired output characteristic than other filter candidates. One filter candidate may be selected that yields an output characteristic. In some examples, the simulation device 100 additionally or alternatively provides, for example, via a screen (not shown), data indicative of the determined output characteristics to the user, and based on the provided data A user may make a selection.

The computer simulations described above may be based on a model of fluid dynamics, for example, using the Navier-Stokes equation to model the flow in the channel and the Brinkman equation for the flow in the membrane 206 . The equations used to model the flow can then be solved to determine the output characteristics. In one example, simulation device 100 may use a finite element method program such as COMSOL® to model the filtration process.

In some examples, the computer simulation process includes a first simulation step comprising determining one or more static process characteristics of the filtration process; and a second simulation stage comprising determining one or more dynamic process characteristics of the filtration process.

Examples of formulas and algorithmic processes that can be used in the simulation process are described in the "Example Algorithms" section below.

The process described above provides an automated method for selecting filters for the filtration process. Prior art methods may involve a process operator to select filters based on their knowledge of filters likely to work for a given filtration process and trial and error using the filters in a test system. However, such methods can be time consuming and expensive. In contrast, the process described above allows the selection of suitable filters for a given filtration process without the time and expense required for testing and without the need for expert knowledge.

Additionally, once a filter is selected, the simulation process described herein may be used to improve or optimize the filtering process performed by the selected filter. For example, once one filter candidate is selected (either automatically or by manual input as described above), further iterations of the simulation process are performed with different process parameters and output according to criteria (which may be set by user input, for example). properties can be improved. For example, the simulator 100 provides an indication of required parameters such as TMP, estimated times to reach desired product concentrations, advisory information such as when to replace filter devices, number of cycles, etc. obtain.

Accordingly, the examples described herein can be used in the design and/or development of filtration processes, eg, in establishing or updating standard operating procedures. In some cases, simulation device 100 may be used in conjunction with an actual (non-simulated) filtration process, which is described below with reference to FIG.

FIG. 5 schematically shows a system including filtration system 500 and simulation device 100 . Filtration system 500 is for performing a filtration process, and in this example, filtration system 500 is for performing a cross-flow filtration process.

Physical filtration system 500 includes supply container 202 and filter device 204 corresponding to the supply container and filter device described above with reference to FIG. A physical filtration system 500 includes a feed pump 501 that flows feed through a filter device 204 via a feed tube 502 . The residue exits the filter via residue tube 504 and returns to container 202 . Filtrate passing through the membrane (not shown) of filter device 204 exits filter device 204 via filtrate tube 506 and is collected in collection tube 508 . Physical filtration system 500 also includes valve 512 that controls flow within system 500 .

According to an example, physical filtration system 500 includes sensors 514a, 514b, 514c that measure one or more characteristics of the filtration process. In the illustrated example, the physical filtration system 500 includes a feed sensor 514a, a retentate sensor 514b, and a filtrate sensor 514c configured to measure one or more characteristics of each of feed, retentate, and filtrate. Properties measured may include, for example, one or more of temperature, pressure, flow rate, and concentration.

Sensors 514a, 514b, 514c are configured to communicate with simulation device 100, for example, via wired or wireless connections.

The simulation device 100 can use data from the sensors 514a, 514b, 514c to change the parameters of the filtration process simulation corresponding to the filtration process being performed by the filtration system 500, which functions as a test system. do. For example, if the measured data received from one or more of the sensors 514a, 514b, 514c indicate characteristic values that deviate from the expected characteristic values determined by the simulation, one or more of the simulation processes can take this into account. parameters can be updated. Optionally, once such a deviation is detected, the filtration process may be repeated (e.g., filter device 204 and feed solution are not by exchanging).

In another example, the simulation device 100 may use data received from the sensors 514a, 514b, 514c to monitor anomalies in the physical filtration system 500, for which an example method of monitoring a filtration process is described. Reference will now be made to FIG. 6, which is a flow diagram illustrating 600.

At 602 , simulation device 100 performs a computer simulation of a filtration process corresponding to the physical filtration process performed by physical filtration system 500 . The simulation may correspond, for example, to a computer simulation as described above. The simulation may be performed based on input data as described above, such as data entered by a user (filter characteristic data, process parameter data, etc.). Additionally or alternatively, simulations may be performed on measurement data received from one or more of sensors 514a, 514b, 514c. For example, sensors provide data indicative of one or more process parameters as described above and form the basis of the simulation.

At 604, the simulation device determines one or more time-varying characteristics of the simulated filtration process based on the simulation process. One or more time-varying characteristics may comprise concentration, eg, filtrate, retentate, or feed concentration or flow rate.

At 606, simulation device 100 receives measurement data from one or more of sensors 514a, 514b, 514c. Measured data indicates the value of one or more time-varying properties at a given time. The given time may be based, for example, on the start time of the filtration process performed by filtration system 500 . For example, the simulation device 100 may receive an indication from the user via the interface 108 at the start of the filtering process, or based on data received from one or more of the sensors 514a, 514b, 514c, for example. time can be detected.

At 608, simulation device 100 compares the measured data received from filtration system 500 to the determined one or more time-varying characteristics. This may comprise, for example, comparing the value of the property indicated by the measured data with the value of the time-varying property determined during the computer simulation process. That is, simulation device 100 may compare measured data values to expected values of the property at a given time as determined by the simulation process.

At 610, simulation device 100 determines whether there is an anomaly in filtration system 500 based on the comparison. For example, the simulation device 100 determines that the filtration system 500 is abnormal when the value indicated by the measurement data differs from the predicted value by more than a predetermined amount (for example, more than a predetermined absolute value or more than a predetermined ratio). can be determined to exist.

If the simulation device 100 determines that an anomaly exists in the filtration system 500 , the simulation device 100 generates an anomaly indicator at 612 . An anomaly indicator may be a visual indicator, such as a warning message or "pop-up" on the screen of simulation device 100, or may be an audible or other warning. In some examples, the anomaly indicator indicates information regarding the nature of the anomaly, eg, anomalous flow rate or concentration of one or more of the feed, filtrate, and residue. In other examples, the anomaly indicator simply indicates the presence of an anomaly without further details.

If the simulation device 100 does not identify an anomaly in the filtration system 500 , the monitoring process returns to 606 to receive more measurement data from the filtration system 500 .

Accordingly, using the method of monitoring the filtration process described with reference to FIG. 6, a user, e.g. Workers can be warned. This allows the user to take corrective action (eg filter replacement) on anomalies and may, for example, improve efficiency or reduce the production of defective products.

Additionally or alternatively, simulation device 100 may store received measurement data and deviations from expected values, e.g., even if the deviation is not large enough to trigger an anomaly indicator. can also be remembered. This data can be used and analyzed, for example, for quality control purposes.

[Algorithm example]
Examples of formulas and algorithms used in the computer simulation process described above are given below. In this example, the first simulation stage uses the following equations:

ここで、
・ ΔＰは、膜横断圧力（ＴＭＰ）、つまり、フィルタ膜２０６にわたる圧力差
・ μは、供給液の粘度
・ Ｒ_ｍは、供給液の粘度と膜の液体透過率に依存する膜起因の抵抗
・ Ｒ_ｇは、ゲルの液体透過率と密度と多孔度に依存する膜２０６起因の抵抗 (Formula 1)

here,
ΔP is the transmembrane pressure (TMP), the pressure difference across the filter membrane 206 μ is the viscosity of the feed liquid R _m is the resistance due to the membrane which depends on the viscosity of the feed liquid and the liquid permeability of the membrane R _g is the resistance due to membrane 206 which depends on the liquid permeability, density and porosity of the gel

ここで、
・ Ｋは、物質移動係数
・ Ｃ_ｍは、膜２０６界面における供給液の濃度
・ Ｃは、バルク溶液の濃度
・ Ｌ_ｐは、膜の液体透過率
・ Ｒ_ｒは、実際の保持力
・ Ａ_１、Ａ_２、Ａ_３はＲ_ｒに依存する定数であり、実験データに少なくとも部分的に基づいて決定され得る。 (Formula 2)

here,
K is the mass transfer coefficient C _m is the concentration of the feed solution at the membrane 206 interface C is the concentration of the bulk solution L _p is the liquid permeability of the membrane R _r is the actual retention force A ₁ , A ₂ , A ₃ are constants that depend on R _r and can be determined based at least in part on experimental data.

ここで、
・ ｕは、供給液のクロスフロー速度
・ Ｄは、関連する拡散係数
・ Ｉは、以下の式４で定義される積分因子 (Formula 3)

here,
u is the feed cross-flow velocity D is the relevant diffusion coefficient I is the integration factor defined in Equation 4 below

ここで、
・ ηは、式５で与えられる無次元変数
・ λは、式６で与えられる (Formula 4)

here,
η is a dimensionless variable given by Equation 5 λ is given by Equation 6

ここで、
・ Ｒｅは、レイノルズ数
・ Ｓｃは、シュミット数
・ ｄ_ｅは、チャネルの等価直径 (Formula 5)

here,
・ Re is the Reynolds number ・ Sc is the Schmidt number ・ d _e is the equivalent diameter of the channel

ここで、
・ Ｊ^＊は、式７で与えられる長さ平均濾液流束 (Formula 6)

here,
• J ^* is the length-average filtrate flux given by Equation 7

(Formula 7)

ここで、
・ Ｃ_ｐは、濾液の濃度 (Formula 8)

here,
・_Cp is the concentration of the filtrate

(Formula 9)

In the above equation, the values of the following filter characteristics can be given as input data as described above: L _p , L, h. The values of the following process parameters can be given as inputs to the simulation as described above: μ, R _m , R _g , D, d _e . The values of the following parameters (referred to herein as "static process parameters") may be determined according to the algorithm: ΔP, u, C, C _m , C _p , J, K.

In some examples, the above equations are used in an iterative algorithm to determine static process parameters. For example, the value of J ^* can be initialized based on previous simulations using similar filter characteristics and process parameters, or can be set with random values. The value of J is determined based on Equation 7. Based on this, the value of K and then the values of ΔP, u, C, C _m , C _p , J are calculated by numerically solving other equations or by other methods. The values of J calculated based on Equation 7 and other equations are then compared. If the calculated values of J differ by more than a predetermined amount, eg, by more than a predetermined percentage (eg, 0.1%), then the process is repeated using a different value of J ^* . Therefore, the process is repeated until the calculated values of J differ by no more than a predetermined amount. When the values of J differ by less than a predetermined value, the value of J ^* and other static process parameters are set for use in the second simulation phase.

The second simulation phase comprises determining the time-varying output characteristics of the filtration process, such as time-varying transmembrane pressure, feed concentration and volume, filtrate concentration and volume, and the like. For example, the second simulation stage may use an equation such as Equation 10 below.

ここで、
・ ｃ_ｆは、供給液濃度
・ ｖ_ｆは、供給液体積
・ ΔＰは、ＴＭＰ
・ Ｌ_ｐは、膜の液体透過率であって、以下の式１１で与えられる
・ Ｍ_Ａは、溶質の分子量 (Formula 10)

here,
_cf is feed concentration _vf is feed volume ΔP is TMP
L _p is the liquid permeability of the membrane and is given by Equation 11 below M _A is the molecular weight of the solute

ここで、
・ Ｌ_ｐ，ｏは、初期液体透過率
・ Δπ_０は、初期浸透圧
・ Δπは、時間変動浸透圧
・ Ｋ_ａ、Ｋ_ｃは、ケーク形成プロセスに関する定数である。これら定数の初期値は仮定される。シミュレーションの動的部分が進行するにつれて、他のパラメータで最適曲線フィッティングを得るためにこれら定数の値を更新することによって、ケーク形成プロセスと、濾過プロセスに対するその影響とをシミュレーションする。 (Formula 11)

here,
• L _p,o is the initial liquid permeability • Δπ ₀ is the initial osmotic pressure • Δπ is the time-varying osmotic pressure • K _a , K _c are constants related to the cake formation process. Initial values for these constants are assumed. As the dynamic portion of the simulation progresses, the cake formation process and its effect on the filtration process are simulated by updating the values of these constants to obtain an optimal curve fit with other parameters.

Many of the examples above are described in terms of cross-flow filtration processes. However, it should be understood that the invention is equally applicable to other types of filtration processes, such as dead-end filtration.

Features described with respect to any one example may be used alone or in combination with other features described herein, and in any other example or combination of other examples. may also be used in combination with one or more features of Additionally, equivalents and modifications not described above may be employed without departing from the scope of the appended claims.

100 simulation device 102 processor 104 computer readable storage medium 106 computer readable instructions 108 interface 200 filtration system 202 supply vessel 204 filter device 206 membrane 208 filter channel 210 permeate side 212 outflow tube

Claims (11)

A method for a simulation device comprising a processor to select a filter for a filtration process, comprising:
the simulation device receiving input data and identifying one or more filter characteristics for each of a plurality of candidate filters for a filtering process based on the input data;
the simulation device identifying process parameters for the filtration process;
the simulation device performing a computer simulation process of the filtering process for each of the plurality of filter candidates based on the identified process parameters and the identified filter characteristics;
the simulation device determining one or more output characteristics of the filtering process based on the computer simulation process;
the simulation device selecting one filter candidate from the plurality of filter candidates based on the one or more output characteristics ;
The computer simulation process comprises:
a first simulation stage comprising determining one or more static process characteristics of the filtering process based at least in part on an iterative algorithm;
and a second simulation stage comprising determining one or more dynamic process characteristics of the filtration process based at least in part on the determined one or more static process characteristics. wherein the input data comprises a filter identifier for each candidate filter, and the method retrieves data from a data store that stores associations between filter identifiers and corresponding filter characteristics to obtain the plurality of candidate filters. 2. The method of claim 1, comprising determining filter characteristics of . 3. The method of claim 1 or 2 , wherein the one or more output characteristics are derived from the determined one or more dynamic process characteristics. 4. A method according to any one of claims 1 to 3 , wherein selecting said one filter candidate is based on comparing said determined one or more output characteristics with a desired output characteristic. 5. A method according to any preceding claim, comprising providing an indication of the one filter candidate selected. performing one or more iterations of the computer simulation process using one or more filter characteristics of the one selected filter candidate, varying process parameters of the filtering process in each iteration, and comparing output characteristics in each iteration; and optionally comparing an iteration-by-iteration output characteristic to a desired output characteristic and selecting parameters for the filtering process based on the comparison. The method described in section. The one or more filter properties include geometric properties and/or porosity properties, the process parameters include one or more parameters relating to the filtration process feed, and the one or more parameters relating to the filtration process feed 7. A method according to any one of claims 1 to 6 , wherein the one or more parameters comprise one or more of initial volume, initial concentration, temperature, solubility and viscosity. 8. The method of any one of claims 1-7 , wherein the one or more output characteristics comprise one or more of product volume, product composition and product viscosity. 9. The method of any one of claims 1-8 , wherein the computer simulation process comprises simulating a solidified solute layer on a filter. 10. Receiving measurement data from a filtration system performing a filtration process corresponding to the simulated filtration process, and optionally modifying the simulation based on said measurement data. The method described in section. 11. The method of claim 10 , comprising determining whether there is an anomaly in a performed filtering process based on comparing the measured data with data from a physical filtering process.

Images